Research On The Seed Metering Performance Testing System Of Corn Metering Device Based On Machine Vision

At present,the amount of corn planted in my country ranks among the top in the world,and corn occupies an important position in the planting structure of the three staple foods in Heilongjiang Reclamation Area.However,the yield of corn per mu needs to be improved compared with developed countries.As the core component of the planter,the equipment with excellent seeding performance is related to the operating quality of the planter,and is one of the important ways to increase crop yields.Therefore,the detection of the seed metering performance of the seed meter has far-reaching significance for the improvement of corn yield and the improvement and innovation of the seed meter.Aiming at the shortcomings of the traditional field test detection method and the butter solid seeding method,this paper consults the research status of the seed metering performance detection technology at home and abroad,and proposes the construction of a corn metering device based on machine vision.,Evaluate the seeding performance of the corn metering device according to the obtained index parameters of the seeding performance.The specific research work of this paper is as follows:(1)Constructing a test bed for the testing system of corn metering deviceModularized design of the hardware structure of the corn seed metering performance detection system,including the seed metering module,the simulated seed bed module,and the image acquisition and processing module.Among them,using the principle of sanding to improve the deficiencies of the butter-spreading solid seed method;the photographing of corn seeds is completed through an industrial camera;the system uses the RS485 bus structure to complete the communication and transmission of the computer to each module,and realizes the simulation of the seed bed module,The running speed of each execution unit in the seeding module is precisely controlled;the industrial camera is calibrated,and the transformation relationship between image pixels and actual physical size is obtained.(2)Video processing of corn seed meteringResearch on the related algorithms of corn seeding video processing,including key frame extraction,image filtering,image stitching,image binarization,corn seed connected regionidentification mark and feature extraction.Among them,for the key frame extraction,an extraction method of converting the seeding video into a static video frame set and then grouping and checking frame by frame is proposed.The threshold shrinkage method of wavelet transform is selected for the key frame image filtering,and the image has been improved and verified.Denoising;use the correlation phase method to obtain the overlap area between key frames,use the SIFT algorithm to obtain the feature points in the overlap area,the BBF algorithm to search for feature points matching,RANSAC algorithm purification,the weighted image fusion technology makes the overlap area smooth transition and eliminate Splicing seams to complete the splicing of key frames;perform grayscale,binarization,and mathematical morphology processing on the spliced image;use the two-scan method to complete the identification and marking of the connected areas of the image corn kernels,and extract the corn Seed grain area,core point coordinate feature quantity,realize the measurement of corn grain distance in the image.(3)Test and Analysis of Seed Metering Performance Testing System of Corn Metering DeviceBased on the corn metering device performance detection system test bench for related experiments.First,the image mosaic effect is tested and analyzed.The test results are consistent with the actual implantation distribution of corn seeds,which verifies the applicability and reliability of the image processing algorithm.Secondly,the single factor test analysis is carried out on the rotating speed of the seeding disc and the vacuum degree of the suction chamber that affect the operation of the seed metering device to obtain the optimal working parameters.Finally,test the performance detection system of the corn metering device.The comparison and analysis of the obtained corn kernel distance data and the manually measured kernel distance data prove that the system can achieve the detection target.